Processing with a million cores: Applications, Tools and Techniques on the Road to Exascale Computing. 14th Biennial ParCo Conference, ParCo2011

Andrew Brown; Jeffrey Reeve; Stephen Furber; David Lester

doi:10.3233/978-1-61499-041-3-327

Processing with a million cores: Applications, Tools and Techniques on the Road to Exascale Computing. 14th Biennial ParCo Conference, ParCo2011

Andrew Brown, Jeffrey Reeve, Stephen Furber, David Lester

Advanced Processor Technology

Research output: Chapter in Book/Report/Conference proceeding › Chapter

Abstract

Traditionally, the performance-limiting attributes of a multi-processor machine are memory bandwidth, the need to maintain state coherence, and simple synchronisation issues, or some combination of these. As the system size increases, so too does the relative cost overhead of solving these problems. The SpiNNaker engine is a million core system with over eight terabytes of distributed memory whose fundamental design axioms neatly sidestep these problems: the processors are not synchronised, the memory is not coherent, and the inter-processor communication is non-deterministic. How, then, can we perform meaningful computations with this architecture? © 2012 The authors and IOS Press. All rights reserved.

Original language	English
Title of host publication	Advances in Parallel Computing\|Adv. Parallel Comput.
Subtitle of host publication	Applications, Tools and Techniques on the Road to Exascale Computing
Place of Publication	Amsterdam, NL,
Publisher	Elsevier BV
Pages	327-334
Number of pages	7
Volume	22
DOIs	https://doi.org/10.3233/978-1-61499-041-3-327
Publication status	Published - 2012

Keywords

discrete event simulation
Massively parallel architecture
multi-core
neural simulation
particle in field

Access to Document

10.3233/978-1-61499-041-3-327

Cite this

Brown, A., Reeve, J., Furber, S., & Lester, D. (2012). Processing with a million cores: Applications, Tools and Techniques on the Road to Exascale Computing. 14th Biennial ParCo Conference, ParCo2011. In Advances in Parallel Computing|Adv. Parallel Comput.: Applications, Tools and Techniques on the Road to Exascale Computing (Vol. 22, pp. 327-334). Elsevier BV. https://doi.org/10.3233/978-1-61499-041-3-327

Brown, Andrew ; Reeve, Jeffrey ; Furber, Stephen et al. / Processing with a million cores : Applications, Tools and Techniques on the Road to Exascale Computing. 14th Biennial ParCo Conference, ParCo2011. Advances in Parallel Computing|Adv. Parallel Comput.: Applications, Tools and Techniques on the Road to Exascale Computing. Vol. 22 Amsterdam, NL, : Elsevier BV, 2012. pp. 327-334

@inbook{907a9491315a41f39cc8bbe5cda976fd,

title = "Processing with a million cores: Applications, Tools and Techniques on the Road to Exascale Computing. 14th Biennial ParCo Conference, ParCo2011",

abstract = "Traditionally, the performance-limiting attributes of a multi-processor machine are memory bandwidth, the need to maintain state coherence, and simple synchronisation issues, or some combination of these. As the system size increases, so too does the relative cost overhead of solving these problems. The SpiNNaker engine is a million core system with over eight terabytes of distributed memory whose fundamental design axioms neatly sidestep these problems: the processors are not synchronised, the memory is not coherent, and the inter-processor communication is non-deterministic. How, then, can we perform meaningful computations with this architecture? {\textcopyright} 2012 The authors and IOS Press. All rights reserved.",

keywords = "discrete event simulation, Massively parallel architecture, multi-core, neural simulation, particle in field",

author = "Andrew Brown and Jeffrey Reeve and Stephen Furber and David Lester",

year = "2012",

doi = "10.3233/978-1-61499-041-3-327",

language = "English",

volume = "22",

pages = "327--334",

booktitle = "Advances in Parallel Computing|Adv. Parallel Comput.",

publisher = "Elsevier BV",

address = "Netherlands",

}

Brown, A, Reeve, J, Furber, S & Lester, D 2012, Processing with a million cores: Applications, Tools and Techniques on the Road to Exascale Computing. 14th Biennial ParCo Conference, ParCo2011. in Advances in Parallel Computing|Adv. Parallel Comput.: Applications, Tools and Techniques on the Road to Exascale Computing. vol. 22, Elsevier BV, Amsterdam, NL, pp. 327-334. https://doi.org/10.3233/978-1-61499-041-3-327

Processing with a million cores: Applications, Tools and Techniques on the Road to Exascale Computing. 14th Biennial ParCo Conference, ParCo2011. / Brown, Andrew; Reeve, Jeffrey; Furber, Stephen et al.
Advances in Parallel Computing|Adv. Parallel Comput.: Applications, Tools and Techniques on the Road to Exascale Computing. Vol. 22 Amsterdam, NL,: Elsevier BV, 2012. p. 327-334.

Research output: Chapter in Book/Report/Conference proceeding › Chapter

TY - CHAP

T1 - Processing with a million cores

T2 - Applications, Tools and Techniques on the Road to Exascale Computing. 14th Biennial ParCo Conference, ParCo2011

AU - Brown, Andrew

AU - Reeve, Jeffrey

AU - Furber, Stephen

AU - Lester, David

PY - 2012

Y1 - 2012

N2 - Traditionally, the performance-limiting attributes of a multi-processor machine are memory bandwidth, the need to maintain state coherence, and simple synchronisation issues, or some combination of these. As the system size increases, so too does the relative cost overhead of solving these problems. The SpiNNaker engine is a million core system with over eight terabytes of distributed memory whose fundamental design axioms neatly sidestep these problems: the processors are not synchronised, the memory is not coherent, and the inter-processor communication is non-deterministic. How, then, can we perform meaningful computations with this architecture? © 2012 The authors and IOS Press. All rights reserved.

AB - Traditionally, the performance-limiting attributes of a multi-processor machine are memory bandwidth, the need to maintain state coherence, and simple synchronisation issues, or some combination of these. As the system size increases, so too does the relative cost overhead of solving these problems. The SpiNNaker engine is a million core system with over eight terabytes of distributed memory whose fundamental design axioms neatly sidestep these problems: the processors are not synchronised, the memory is not coherent, and the inter-processor communication is non-deterministic. How, then, can we perform meaningful computations with this architecture? © 2012 The authors and IOS Press. All rights reserved.

KW - discrete event simulation

KW - Massively parallel architecture

KW - multi-core

KW - neural simulation

KW - particle in field

U2 - 10.3233/978-1-61499-041-3-327

DO - 10.3233/978-1-61499-041-3-327

M3 - Chapter

VL - 22

SP - 327

EP - 334

BT - Advances in Parallel Computing|Adv. Parallel Comput.

PB - Elsevier BV

CY - Amsterdam, NL,

ER -

Brown A, Reeve J, Furber S, Lester D. Processing with a million cores: Applications, Tools and Techniques on the Road to Exascale Computing. 14th Biennial ParCo Conference, ParCo2011. In Advances in Parallel Computing|Adv. Parallel Comput.: Applications, Tools and Techniques on the Road to Exascale Computing. Vol. 22. Amsterdam, NL,: Elsevier BV. 2012. p. 327-334 doi: 10.3233/978-1-61499-041-3-327

Processing with a million cores: Applications, Tools and Techniques on the Road to Exascale Computing. 14th Biennial ParCo Conference, ParCo2011

Abstract

Keywords

Access to Document

Fingerprint

Cite this